最优化牛顿法最速下降法共轭梯度法matlab代码(共8页).docx

精选优质文档-倾情为你奉上牛顿法迭代公式：Matlab代码：function x1,k =newton(x1,eps)hs=inline(x-1)4+y2); 写入函数ezcontour(hs,-10 10 -10 10); 建立坐标系hold on; 显示图像syms x y 定义变量f=(x-1)4+y2; 定义函数grad1=jacobian(f,x,y); 求f的一阶梯度grad2=jacobian(grad1,x,y); 求f的二阶梯度k=0; 迭代初始值while 1 循环grad1z=subs(subs(grad1,x,x1(1),y,x1(2); 给f一阶梯度赋初值grad2z=subs(subs(grad2,x,x1(1),y,x1(2); 给f二阶梯度赋初值 x2=x1-inv(grad2z)*(grad1z); 核心迭代公式if norm(x1-x2)eps 判断收敛条件 break;else plot(x1(1)